Analisis Data Kuanti-mpkt Mba Titut

  • Uploaded by: Communication Management UI
  • 0
  • 0
  • November 2019
  • PDF

This document was uploaded by user and they confirmed that they have the permission to share it. If you are author or own the copyright of this book, please report to us by using this DMCA report form. Report DMCA


Overview

Download & View Analisis Data Kuanti-mpkt Mba Titut as PDF for free.

More details

  • Words: 2,784
  • Pages: 75
Analisis Data Univariat, Bivariat, Multivariat

Beberapa faktor yang mempengaruhi analisis data • Populasi – ukuran populasi (parametrik: (e.g., µ, σ, σ2 , ρ ) atau non parametrik • Penarikan sample – probabilita atau non probabilita • Jenis variabel – diskret atau continue, • Skala pengukuran – nominal, ordinal, interval atau rasio • Hasil kurva normal – normal atau tidak

Analisis berdasar jumlah variabel • Univariate Statistics melibatkan satu variabel pada satu waktu • Bivariate Statistics melibatkan dua variabel secara simultan • Multivariate Statistics melibatkan tiga atau lebih variabel dalam satu analisis

Klasifikasi Tehnik Statistik Univariat (inferensi) Univariate Techniques

Metric Data (i.e., interval or ratio)

One sample -t-test, z- test

Independent t-test z-test One-way ANOVA

Two or more samples

Related Paired t-test

Non-metric data (i.e., nominal, ordinal)

One sample Frequency, Chisquare, K-S, etc.

Independent Chi-square Mann-Whitney K-S, etc.

Two or more samples

Related Wilcoxon McNemar Chi-square, etc.

Univariat analisis – Metrik data satu sampel • Satu sample - t test (sampel kecil), z test (sampel besar) Suatu pengujian dilakukan kepada 12 murid SMP. Ratarata tingkat melek media mereka adalah 73,8 dg standard deviasi 7,9. Seorang peneliti menarik kesimpulan: jika rata-rata melek media mereka kurang dari 75, maka mereka diharuskan mengikuti kursus melek media. Ujilah pendapat peneliti dengan tingkat signifikansi 95%.

• sehingga dapat disimpulkan bahwa dengan Ho tidak ditolak, artinya tidak cukup bukti untuk menyatakan bahwa rata-rata melek media kurang dari 75, sehingga murid SMP tidak perlu kursus melek media t hitung

73,8 − 75 = = −0.53 7,90 12 Daerah Ho ditolak

Univariat data – metrik, dua sampel atau lebih • Independent: melihat perbedaan jika dua kelompok sampel diteliti, namun tidak terdapat hubungan di antara kedua kelompok tersebut • Related/dependent: melihat perbedaan jika dilakukan dua kali pengujian untuk kelompok yang sama pada waktu yang berbeda • One way ANOVA melibatkan hanya satu variabel kategorikal, bertujuan untuk menguji

hipotesis nol yang artinya, rata-rata kategori sama dengan rata-rata di populasi Ho: μ1=μ2=μ3...=μn

Univariat analisis – non metrik satu sampel • Tabel distribusi frekuensi • Chi square test for goodness of fit: untuk mengetahui perbedaan proporsi dari masing-masing kategori digunakan dengan anggapan distribusi data tidak simetris Dengan rumusan hipotesis sebagai berikut: Ho: p1 = p2 = pn = 1/jumlah kategori Ha: p1 = p2 ≠ pn = 1/jumlah kategori

• Lakukan pengujian dengan tingkat signifikansi 95% untuk hasil penelitian mengenai afiliasi masyarakat pada partai politik tertentu (data fiktif) apakah proporsinya berbeda antara kategori yang satu dengan kategori yang lainnya. • sehingga dapat disimpulkan bahwa dengan tingkat signifikansi 95% cukup bukti untuk menyatakan bahwa proporsi antara kategori yang satu dengan yang lainnya tidak sama dengan 1/5.

Afiliasi masyarakat pada parpol

f

PK

31

PAN

19

PBB

19

P. Golkar

25

PDIP

41

Jumlah

135

Daerah Ho ditolak

9,49

12,74

Bivariate Analysis • Describes the association/correlation between each pair of variables • Answers the question: Is there a relationship between these two variables? • Initial step in hypothesis testing – melibatkan signifikansi (significance of null hypothesis)

Beberapa formulasi hub. bivariat • Cross sectional - concomitant variaton (pearson’s r) - differentiation (t-test, independent sample) - probability distribution (cramer, lambda, spearman, kendall) • Longitudinal - before – after differentiation (t-test dependent)

BIVARIAT - KORELASI Variabel 1 Nominal nominal

Variabel 2 Ordinal

Chi-square Cramer’s V Coefficien contingency

Interval T test Z test

Lambda

ordinal

interval

Kendall’s Spearman Gamma Sommer’s D Pearson’s Regression

NOMINAL

The Phi Coefficient (φ) • Based on chi-square • Tehnik utk melihat keeratan 2 var. nominal, atau var. yg tlh diredusi jd skala nominal • Data disajikan dalam tabel silang 2x2

Interpretation There is a significant association between race and sentence type in the population of the order of 0.149.

Contingency Coefficient (C) • Keeratan antara dua nominal variabel pada tabel frekuensi 2x2 ATAU LEBIH • Dua variabel dari skala nominal atau variabel yang telah diredusi ke dalam skala nominal

Level of significance: p < 0.001 Contingency Coefficient = 0.324

Cramér’s V Correlation Coefficient • Menjelaskan korelasi antar dua var. nominal • Alternatif lain penggunaan Contingency Coefficient C bila data pada tabel frekuensi tidak simetris (cth: 2x3, 4x7, 5x6)

Lambda (λ): Guttman’s Coefficient of Predictability • Tehnik untuk menentukan tingkat kesalahan prediksi pada satu var. nominal dapat dikurangi dengan mengetahui var. nominal lainnya.

• Interpretation The error in predicting gender is reduced by 0.21, (21%), by knowledge of attitude toward the death penalty. Since λ = 0.0, the reduction in the error in predicting attitude toward the death penalty by a knowledge of a person’s gender would be 0%, or none at all.

ORDINAL

Spearman Rank-Order Correlation Coefficient: ρ (rho) • Untuk dua variabel ordinal atau variabel metrik yg diredusi ke dalam skala ordinal • Korelasi antar variabel bersifat linear • Bila uji signifikansi digunakan, sampel diambil secara random dr populasi

Goodman’s & Kruskal’s Gamma (γ) • Menjelaskan keeratan antar dua var. ordinal • Hubungan antar var. linear

INTERVAL

Pearson’s Correlation Coefficient (r) • Tell you two things about the relationship: – Strength? – Direction?

• Also, look at the p-value: – Significant?

Strength • How strong is the relationship? • Look at the value of r • How big is the number? – – – – –

1.0 (-1.0) = Perfect Correlation .60 to .99 (-.60 to -.99) = Strong .30 to .59 (-.30 to -.59) = Moderate .01 to .29 (-.01 to -.29) = Weak 0 = No Correlation

Direction • What is the direction of the relationship? • Look at the sign of r • Positive (+) – Both variables move in the same direction – If one is going up, the other is going up too. – OR, if one is going down, the other is going down too.

• Negative (-) – Both variables move in opposite directions – If one is going up, the other is going down. – OR, if one is going down, the other is going up.

Significance • Is the relationship significant? – Look at the p-value • p < .05, then it is significant • p > .05, then it is NOT significant

• NOTE: You may see the strength and direction of a relationship, but it may not actually be significant.

REGRESSION ANALYSIS • Melakukan prediksi terhadap variabel dependen berdasarkan variabel independen

Cross Tabulation Count

Absenteesim Total

no yes

Satisfaction no yes 5 9 4 2 9 11

Total 14 6 20

BIVARIAT – PERBEDAAN RATA-RATA INDEPENDENT 2 SAMPEL

>2 SAMPEL

• Binomial Test • Kruskal• Mann-Whitney Wallis H U Test Test • KolmogorovSmirnov Z test • WaldWolfowitz runs test • Mosesextremes reaction

PAIRED 2 SAMPEL

>2 SAMPEL

• Wilcoxon Test • Sign Test • Mc-Nemar Test

• Friedman Test • Kendall’s W Test • Cochran’s Q Test

BINOMIAL TEST  Untuk menguji perbedaan proporsi antara 2 kelompok  Dua kelompok di uji tentang keberhasilan sosialisasi perda no. 5 tahun 2004. Kelompok 1: dari 60 responden, 41 menyatakan setuju. Kelompok 2: dari 60 responden, 24 menyatakan setuju  Apakah ada perbedaan proporsi yang setuju antara kelompok 1 dengan kelompok 2 dengan α=5%

Ho: Tidak ada perbedaan proporsi jumlah yang setuju antara kelompok 1 dengan kelompok 2 Ha: ada perbedaan proporsi jumlah yang setuju antara kelompok 1 dengan kelompok 2

MANN-WHITNEY U TEST  Untuk menguji ada tidaknya perbedaan skor antara dua kelompok yang independen

Contoh Mann-Whitney U Test • Penelitian dilakukan untuk menguji perbedaan skor partisipasi murid sekolah agama dan sekolah umum. Hasilnya:

Sekolah Agama 5 11 19 23 6 13 19 24 8 13 20 28 8 14 21 28 10 16 22 10 17 22 11 17 22

Sekolah Umum 5 13 19 26 7 13 20 26 8 14 20 27 8 14 21 27 8 18 22 9 19 24 12 19 24

Ranks partisipa

sekolah Sekolah Agama Sekolah Umum Total

N

Te st Statisticsa Mann-Whitney U Wilcoxon W Z Asymp. Sig. (2-tailed)

partisipa 296.500 621.500 -.311 .756

a. Grouping Variable: sekolah

25 25 50

Mean Rank 24.86 26.14

Sum of Ranks 621.50 653.50

UJI PERINGKAT KRUSKALWALLIS • Digunakan untuk skala ordinal 3 kelompok atau lebih sampel independen • Data tidak terdistribusi secara normal dan variansi populasi tidak sama

Ho: tidak ada perbedaan skor antara kelompok industri, keuangan dan perdagangan Ha: ada perbedaan skor antara kelompok industri, keuangan dan perdagangan Industri 32(2) 30(4,5) 30(4,5) 29(6) 26(9) 23(11) 20(14,5) 19(16,5) 18(18) 12(27) n1=10 ΣR1=113

Keuangan 32(2) 32(2) 26(9) 26(9) 22(12) 20(14,5) 19(16,5) 16(19) 14(24) 14(24) n2=10 ΣR2=132

Perdagangan 28(7) 21(13) 15(20,5) 15(20,5) 14(24) 14(24) 14(24) 11(28) 9(29) 8(30) n3=10 ΣR3=220

Ranks skor

bidang industri keuangan perdagangan Total

N 10 10 10 30

Te st Statisticsa,b Chi-Square df Asymp. Sig.

skor 8.471 2 .014

a. Kruskal Wallis Test b. Grouping Variable: bidang

Mean Rank 19.70 17.80 9.00

UJI PERINGKAT BERGANDA WILCOXON • Uji perbedaan skor sebelum dan sesudah

produksi Operator sebelum sesudah A 17 18 B 21 23 C 25 22 D 15 25 E 10 28 F 16 16 G 10 22 H 20 19 I 17 20 J 24 30 K 23 26

d 1 2 -3 10 18 0 12 -1 3 6 3

ranking tanda + 1,5 1,5 3 3 5 5 8 8 10 10 9 1,5 5 7 5

9 1,5 5 7 5

Ho: tidak ada perbedaan kualutas produksi sebelum dan sesudah penggunaan mesin baru Ha: ada perbedaan kualutas produksi sebelum dan sesudah penggunaan mesin baru

Ranks N after - before Negative Ranks Positive Ranks Ties Total

2a 8b 1c 11

Mean Rank 3.25 6.06

a. after < before b. after > before c. after = before

Te st Statisticsb after - before Z -2.148 a Asymp. Sig. (2-tailed) .032 a. Based on negative ranks. b. Wilcoxon Signed Ranks Test

Sum of Ranks 6.50 48.50

McNEMAR TEST  Uji perubahan sikap sebelum dan sesudah

After Before

Success

Failure

Failure

n1

n2

Success

n3

n4

postconv preconv 1

2

1

15

0

2

15

20

Te st Statisticsb preconv & postconv N 50 Exact Sig. (2-tailed) .000 a a. Binomial distribution used. b. McNemar Test

FRIEDMAN TEST Digunakan untuk melihat ada tidaknya perubahan hasil untuk pengujian > 2 kali.

Ho: tidak ada perubahan skor hasil tes Ha: ada perubahan skor hasil tes

(

)

12 2 2 2 χ = 19 + 29 + 42 − ( 3(15)(3 + 1) ) = 17,73 15(3)(3 + 1) 2 F

Bandingkan dengan nilai tabel chi-square pada df=k-1 yaitu 5,99 Jika uji Friedman > Chi-square tabel maka Ho ditolak

Ranks tes1 tes2 tes3

Mean Rank 1.27 1.93 2.80

Te st Statisticsa N Chi-Square df Asymp. Sig.

15 18.345 2 .000

a. Friedman Test

COCHRAN’S TEST Uji perbedaan sikap pada > 2 hal yang kategori jawabannya hanya 2.

Ho: tidak ada perbedaan jawaban untuk ketiga jenis wawancara Ha: ada perbedaan jawaban untuk ketiga jenis wawancara

[

]

(3 − 1) (3(52 + 52 + 0 2 ) − 10 2 Q= = 7,143 3(10) − 16 Bandingkan nilai Cochran’s Q dengan nilai tabel chi-square pada df=k-1 yaitu 5,99. Hasil Q > dari nilai chi-square sehingga Ho ditolak artinya ada perbedaan jawaban ya untuk ketiga jenis wawancara.

Fre que ncie s Value 0 VAR00023 VAR00024 VAR00025

1 5 5 10

Te st Statistics N Cochran's Q df Asymp. Sig.

10 7.143 a 2 .028

a. 0 is treated as a success.

5 5 0

ELABORASI • Merinci penjelasan hubungan antar variabel – dengan memasukkan variabel ketiga dalam analisis • Variabel ketiga: intervening variable(s), antecedent variable(s), specifying variable(s)

Status Analisis dalam Elaborasi •Eksplanasi theoritical status Tipe brand placement (antecedent variable) – Terpaan Media (independent variable) – Interaksi parasosial (intervening variable) – perilaku membeli (dependent variable) •Prediksi Theoritical status Faktor-faktor yang mempengaruhi perilaku membeli

Intervening Variable(s) • All three variables (intervening, independent, dependent) must be related • When intervening variable is controlled, the relationship between IV and DV should vanish • When IV is controlled, the relationship between the intervening and DV should not disappear • Antecedent – independent – intervening dependent

ANTECEDENT VARIABLE(S) • All three variables (antecedent, independent and dependent) must be related • When antecedent variable is controlled, the relationship between IV and DV should vanish • When IV is controlled, the relationship between the antecedent and DV should not disappear

Kemungkinan hasil elaborasi • • • •

Konstan - replikasi Melemah – eksplanasi/interpretasi Terbelah – spesifikasi Menguat – variabel ketiga mempengaruhi (suppressor)

Klasifikasi tehnik statistik multivariat Multivariate Techniques

Dependence Techniques

One Dependent Variable -Cross Tabulation - ANOVA - ANCOVA - Multiple Regression - Discriminant Analysis - Conjoint Analysis

More than one Dependent Variables - Multivariate analysis of variance and covariance - Canonical correlation - Multiple discriminant analysis

Interdependence Techniques

Variable interdependence - Factor Analysis

Inter-object similarity - Cluster Analysis - Multidimensional Scaling

MULTIPLE REGRESSION (REGRESI BERGANDA)

Perbedaan MR dan tehnik lain • Satu DV (criterion variable) & beberapa IV (predictor variable) (bandingkan dg factor analysis & cluster analysis yg memilik banyak DV) • IVs boleh saling berhubungan (bandingkan dg ANOVA yg IVs-nya harus saling independen) • IVs karakternya bisa kontinu (bandingkan dg ANOVA yg IVs-nya harus kategorial)

• Pengamatan thd satu DV dijelaskan melalui variasi hubungannya dg IVs • Mengetahui mana dari IV yang secara signifikan berkorelasi dg DV, dg mempertimbangan bermacam korelasi yang muncul di antar IVs • Hubungan antarvariabel dalam MR bersifat alamiah, bukan eksperimental

Rules in Multiple Regression • Random sampling • DV: interval (continuous scale), tp praktiknya skala ordinal sdh cukup (kecuali jika n kecil sekali) • IV: interval, but most ordinal scale measurement will be acceptable in practice • Normal distribution • Relationship between IV and DV should be linear. • Although IV can be correlated, there must be no perfect correlations among them (correlation: the score on one variable will allow you to predict the score on the other variable) • Tdk boleh ada interaksi antar IVs – pengaruh tiap IV thd DV harus independen satu sama lain

Kapan menggunakan MR? • when exploring linear relationships between the predictor and criterion variables – that is, when the relationship follows a straight line. • The criterion variable should be measured on a continuous scale (such as interval or ratio scale). • The predictor variables should be measured on a ratio, interval, or ordinal scale. • Multiple regression requires a large number of observations (N) (Jmlh N harus melebihi IVs) • Minimun: N= 5 x IVs • Medium: N = 10 x IVs • Ideal: N = 40 x IVs

Path Analysis

Feature of Path Analysis • Path analysis is an extension of the regression model • Used to test the fit of the correlation matrix against two or more causal models which are being compared by the researcher. • A regression is done for each variable in the model as a dependent on others which the model indicates are causes. • The regression weights predicted by the model are compared with the observed correlation matrix for the variables, and a goodness-of-fit statistic is calculated. • The best-fitting of two or more models is selected by the researcher as the best model for advancement of theory.

CLUSTER ANALYSIS

Features of Cluster Analysis • Descriptive, exploratory technique • Non-inferential: no statistical inference from sample to population • Wide variety of methods: can produce many different “solutions” • Clusters always produced, regardless of “true structure” • Highly dependent on the variables used

TUJUAN DASAR • Mengelompokkan objek berdasarkan kesamaan karakteristik • Objek yang sama akan berada dalam satu cluster

Tujuan Partition cases into groups based on similarities of characteristics: •To form a taxonomy (empirically-based classification) •To compare with a typology (theoretically based classification) •To simplify data structure (clusters can be profiled by their general characteristics) •To reveal hidden relationships among cases

DISCRIMINANT ANALYSIS

Two goals for DA • Interpretation: “How are the groups different?” Find and interpret linear combinations of variables that optimally predict group differences (to assess the adequacy of classification, given the group memberships of the objects under study ) • Classification: “How accurately can observations be classified into groups?” Using functions of variables to predict group membership for a data set and evaluate expected error rates (to assign objects to one of a number of (known) groups of objects)

Discriminant Analysis: Introduction • A technique use to build a predictive model of group membership based on observed characteristics of each case (e.g. to group children into two main groups of Very Clever or Just Clever children based on their performance on the three core subjects English, Mathematics, and Science. )

• to generate functions from a sample of cases for which group membership is known; the functions can then be applied to new cases with measurements for the predictor variables but unknown group membership (e.g. That is, knowing a child's score on three subjects, we can use the discriminant function to determine whether the child belongs to the Very Clever group or the Just Clever group. )

Related Documents

Analisis Data
May 2020 39
Analisis Data
December 2019 49
Analisis Data
August 2019 75

More Documents from "Noval Bastian"

Varel2001
November 2019 52
Operasionalisasi Konsep
November 2019 51
Budget
October 2019 64
Tingkatan Komunikasi
November 2019 48